Related Experiment Video
Updated: May 27, 2026
07:05
Microwave-Assisted Extraction of Phenolic Compounds and Antioxidants for Cosmetic Applications Using Polyol-Based Technology
Published on: August 23, 2024
Ensemble machine learning with Bayesian optimization predicts bioactive extraction from fermented watermelon rind
Mostafa Khajeh1,2, Mansour Ghaffari-Moghaddam3,4, Afsaneh Barkhordar3
1Department of Chemistry, Faculty of Science, University of Zabol, Zabol, Iran. m_khajeh@uoz.ac.ir.
Scientific Reports
|May 25, 2026
Summary
Solid-state fermentation and green extraction methods significantly boost bioactive compound yields from watermelon rind. Machine learning accurately predicts and optimizes extraction conditions for nutraceutical applications.
Area of Science:
- Agricultural Science
- Food Science
- Biotechnology
Background:
- Watermelon rind is an agricultural waste rich in bioactive compounds.
- Traditional extraction methods can be inefficient and environmentally taxing.
- Sustainable valorization of agricultural byproducts is crucial for the circular economy.
Purpose of the Study:
- To develop an integrated, green extraction strategy for bioactive compounds from fermented watermelon rind.
- To optimize extraction parameters using machine learning.
- To evaluate the efficacy of solid-state fermentation as a pretreatment method.
Main Methods:
- Combined natural deep eutectic solvents (NADES), microwave-assisted extraction (MAE), and machine learning (Bayesian optimization).
- Utilized solid-state fermentation pretreatment.
- Developed ensemble models to predict total phenolic content (TPC), total flavonoid content (TFC), and antioxidant activity (DPPH).
Main Results:
- Solid-state fermentation significantly increased extraction yields of phenolics, flavonoids, and antioxidant activity.
- Ensemble models demonstrated high predictive accuracy (R² > 0.91) with minimal overfitting.
- Identified temperature and solid-liquid ratio as key extraction parameters.
Conclusions:
- The integrated approach of fermentation, NADES, MAE, and machine learning offers a sustainable and efficient method for extracting bioactive compounds.
- This strategy enables the valorization of watermelon rind for nutraceutical applications.
- The developed models accurately predict and optimize extraction, showing potential for industrial scalability.
